[Bug c++/77563] [5/6/7 Regression] explicit constructor breaks narrowing conversion overload resolution

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=77563

Jakub Jelinek <jakub at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |jakub at gcc dot gnu.org,
                   |                            |jason at gcc dot gnu.org
   Target Milestone|---                         |5.5

--- Comment #4 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
The 4.5 -> 4.6 difference appeared in r159332:

./cc1plus.159325 -quiet -std=c++0x pr77563.C
pr77563.C: In function ‘int main()’:
pr77563.C:12:7: error: conversion from ‘long int’ to ‘A’ is ambiguous
pr77563.C:3:3: note: candidates are: A::A(unsigned int)
pr77563.C:2:3: note:                 A::A(int)
./cc1plus.159332 -quiet -std=c++0x pr77563.C
pr77563.C: In function ‘int main()’:
pr77563.C:12:7: error: converting to ‘A’ from initializer list would use
explicit constructor ‘A::A(long int)’

then r229283 started to ICE on it instead, and finally r236395 fixed the ICE,
but emitted a note without corresponding error/warning:

./cc1plus.236395 -quiet -std=c++0x pr77563.C
pr77563.C: In function ‘int main()’:
pr77563.C:8:6: note:   initializing argument 1 of ‘void f(A)’
 void f(A) { }
      ^